PhD, quantitative research. Public repositories below: identification and statistics on simulated designs, forecasting and optimisation notes, valuation identities, structured reviews of quantitative answers, and Lean 4 + mathlib studies.

Samples in the Python labs are simulated DGPs or closed forms unless a file says otherwise. Recovering a simulated parameter is not an empirical finding. Nothing here is a desk, a trading book, or a commercial deployment. Lean files that compile are theorems about the statements as written.

econometrics-causal-inference-lab — DiD, IV, RD, matching, and panel methods on documented DGPs. A coefficient that prints is not an identified treatment effect. CASE_STUDY.md

statistical-reasoning-validation — Type I / Type II, coverage, p-value misuse, omitted-variable bias, under known DGPs. statistical_error_catalogue.md

A longer list of those identification and inference labs: ECONOMICS_EXPERT_PORTFOLIO.md.

time-series-forecasting-lab — walk-forward skill on simulated series, not in-sample fit.

optimization-decision-models — write the programme, solve it, substitute the point back into the original functions. Solver success is not KKT.

quantitative-finance-models — TVM, duration, parity, VaR/ES identities under stated assumptions. Not a trading book.

These are method laboratories, not an engineering job title.

computational-ml-stem-problem-forge — problems with a reference solver and two further checks that are not copies of that solver.

machine-learning-model-selection-lab — invalid workflows kept next to the matching design (full-frame scaling, group leakage, inner best_score_ treated as generalisation).

ai-response-evaluation-benchmarks — fluent answers that fail on the target, the information set, or the interpretation. One author coded the YAML.

genai-rag-evaluation-lab — gold can sit at rank 1 while the extractive answer abstains. Retrieval metrics are not answer quality.

sql-ml-feature-engineering-lab — a join without txn_ts <= cutoff admits a planted future value without selecting the label.

pytorch-deep-learning-lab — hand derivatives, finite differences, and autograd; they disagree on purpose at a ReLU kink.

mlops-reproducible-serving-lab — HTTP 200 with the wrong probability after a column swap.

Notes on those labs: MACHINE_LEARNING_EXPERT_PORTFOLIO.md.

Self-directed 10-option items (one key, nine distractors, a derivation, references) in microstructure, macroprudential policy, behavioral finance, urban economics, and tokenomics. Not a venue, central-bank, or protocol job.

economics-finance-assessment-benchmark-lab — accepted items and rejected drafts. CI checks schema, not that the economics is true.

Companions: microstructure · macroprudential · behavioral finance · urban · tokenomics.

Map: ECONOMICS_FINANCE_ASSESSMENT_PORTFOLIO.md.

Compilation is not faithfulness.

• lean4-optimization-economics — small programmes and budget sets
• lean4-analysis-formalization — limits, continuity, compactness
• lean4-mean-value-theorems — Rolle, Lagrange, Cauchy reconstructed from EVT and Fermat
• lean4-formalization-faithfulness — compiled theorems that still miss the intended claim
• lean4-formalization-review — review cases
• lean4-proof-engineering — structured proofs and tactic habits
• lean4-automation-debugging — failed proofs, classified and repaired

Python, R, SQL, NumPy, Pandas, SciPy, statsmodels, scikit-learn, PyTorch, Lean 4/mathlib.